The lemon hypothesis vs evidence

Back in 1970 George Akerlof wrote “The Market for Lemons”, where he described a game where if, buyers have less information than sellers, it is possible that mutually beneficial trade may not occur. (Wikipedia)

Now I’ve noticed a bunch of my favourite economics blogs discussing this paper by Arif Sultan, on empirical evidence and the “lemons problem”. (Blogs are Anti-Dismal, Division of Labour, and Marginal Revolution). The paper appears to say that, empirically, the quality of new and used cars is the same – something that would not occur in the case of a “market failure” based on asymmetric information.

Overall I enjoyed the posts offered by Anti-Dismal (descriptive) and the Division of Labour (stating it was an interesting result), but I think Marginal Revolution takes more out of the paper than it actually offers. I can’t actually get access to the paper – so I’m having to work off the abstract here.

On Marginal Revolution, Tyler Cowen states that this paper provides more evidence against the lemons model. The paper definitely shows that the market that is analysed doesn’t suffer from the market failure that can occur in the case of the lemons problem – however, it doesn’t show that the lemons phenomenon doesn’t hold true in the car market, and even if it did it also doesn’t negate the usefulness general lemons model in anyway (which is too do with asymmetric information on the buyers side of a market, not just used cars).

But this problem can’t exist if the market is functioning!

Not so. In the lemons model the buyer does not know the quality of an individual vehicle. However, they do know the distribution of vehicles in the market. Buyers then base the price they are willing to pay for a vehicle on an expectation of the quality of the vehicle, which depends on this distribution.

Each car type will sell as long as this price is greater than the price the seller would accept. If the sellers of good quality used cars are willing to sell at a sufficiently low price, there is no need for the market failure to occur. As a result, we will have a fully functioning market in “used cars” – however asymmetric information will remain as a defining feature of the market.

As a result, I’m wondering – what is with the hating on the lemon’s model? The result shown by the paper is interesting, but it in no way contradicts the usefulness of the general lemons result – that in the face of asymmetric information on the buyers side, efficient transactions may not occur.

Add to FacebookAdd to DiggAdd to Del.icio.usAdd to StumbleuponAdd to RedditAdd to BlinklistAdd to Ma.gnoliaAdd to TechnoratiAdd to FurlAdd to Newsvine

6 replies
  1. Paul Walker
    Paul Walker says:

    “In the lemons model the buyer does not know the quality of an individual vehicle. However, they do know the distribution of vehicles in the market. Buyers then base the price they are willing to pay for a vehicle on an expectation of the quality of the vehicle, which depends on this distribution.

    Each car type will sell as long as this price is greater than the price the seller would accept. If the sellers of good quality used cars are willing to sell at a sufficiently low price, there is no need for the market failure to occur. As a result, we will have a fully functioning market in “used cars” – however asymmetric information will remain as a defining feature of the market.”

    While this is true it doesn’t explain the no difference in maintenance expenditures result since under the above assumption low quality cars could still have higher maintenance expenditures.

  2. Matt Nolan
    Matt Nolan says:

    Hi Paul, thanks for commenting.

    “While this is true it doesn’t explain the no difference in maintenance expenditures result since under the above assumption low quality cars could still have higher maintenance expenditures.”

    Just a sec, when we are discussing quality I take it that we are implicitly talking about the maintenance expenditure – so I implicitly do assume that the higher the maintenance expenditure the lower the quality.

    As a result, this empirical evidence suggests that used cars are not lower quality than new cars (I believe they adjust for age in this) and as a result, no market failure is prevalent (which isn’t a conclusive result, but is very convincing).

    We could still have a situation where we have asymmetric information and no market failure as long as the sellers reservation value is sufficiently low relative to the buyers willingness to pay for a good of unknown quality.

    To show that this isn’t evidence of the lemons argument being flawed I only have to show that the empirical result is consistent with a plausible set of valuations on the behalf of the buyer and seller – I don’t need to show reverse, that the lemons market will uniquely define an eqm where the maintenance costs are the same.

    As a result, this evidence indicates that there is no market failure – however, it does not indicate a failure of the lemons model, which is the point I was trying to address in the post.

    Has that helped to clarify my point, or do I still sound like I’m talking crack 🙂

  3. sparrow
    sparrow says:

    You should have the decency to at least put the exact link or title of the original Cowen and other blog posts. It is annoying because you didn’t.

Trackbacks & Pingbacks

  1. […] Assets and the market for lemons 1 10 2008 I have spent so much time blabbing about asymmetric information without every explaining what I meant. As a result I feel that everyone deserves a little explanation.  Thanks goes out to Akerlof and the lemons hypothesis. […]

Comments are closed.